The S100A8/A9 protein as a partner for the cytosolic factors
of NADPH oxidase activation in neutrophils
Jacques Doussiere, Farid Bouzidi and Pierre V. Vignais
Laboratoire de Biochimie et Biophysique des Syste
`
mes Inte
´
gre
´
s (UMR 5092 CEA-CNRS-UJF), De
´
partement Re
´
ponse
et Dynamique Cellulaires, CEA-Grenoble, France
In a previous study, the S100A8/A9 protein, a Ca
2+
-and
arachidonic acid-binding protein, abundant in neutrophil
cytosol, was found to potentiate the activation of the redox
component of the O
2
–
generating oxidase in neutrophils,
namely the membrane-bound flavocytochrome b,bythe
cytosolic phox proteins p67phox, p47phox and Rac
(Doussie
`
re J., Bouzidi F. and Vignais P.V. (2001) Biochem.
Biophys. Res. Commun. 285, 1317–1320). This led us to check
by immunoprecipitation and protein fractionation whether

membrane-bound flavocytochrome b.
Keywords: NADPH oxidase activation; superoxide O
2
–
;
neutrophils; phox proteins; S100A8/ A9 protein.
The heterodimeric Ca
2+
-binding protein S100A8/A9, also
referred to in the literature as MRP8/MRP14, is expressed
constitutively in large amounts in neutrophils and mono-
cytes [1–3], where it plays a role in the activation process
(reviewed in [4]) and adhesion [5]. The S100A8/A9 protein
might also serve as a reservoir and a carrier of arachidonic
acid [6–8]. This latter finding is all the more noteworthy as
arachidonic acid is currently used in cell-free system to
activate the O
2
–
generating NADPH oxidase, an enzymatic
complex responsible for the microbicidal function of
neutrophils and macrophages [9]. In its activated form,
the NADPH oxidase complex is composed of a membrane-
bound flavocytochrome b and proteins of cytosolic origin,
called phox (for phagocyte oxidase) factors of oxidase
activation or cytosolic phox proteins, namely p67phox,
p47phox and Rac (reviewed in [10]). An additional cytosolic
protein p40phox has been found to copurify with p47phox
and p67phox and the three cytosolic factors appear to form
an activation complex [11]. The O

E-mail: [email protected]
Abbreviations: PMSF, phenylmethanesulfonyl fluoride.
(Received 14 February 2002, revised 30 April 2002,
accepted 17 May 2002)
Eur. J. Biochem. 269, 3246–3255 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.03002.x
enhanced [16]. Using a PAO affinity chromatography
column, we identified in neutrophil cytosol the S100A8/A9
protein as the PAO target responsible for the increased
oxidase activation [15]. Here we report immunoprecipita-
tion and protein fractionation experiments that suggest that
the S100A8/A9 protein interacts with the cytosolic factors
of oxidase activation and more preferentially with p67phox.
We also report a study, in a cell-free system, of the effects of
S100A8/A9 on the kinetics of oxidase activation.
MATERIALS AND METHODS
Chemicals
NADPH, GTPcS, leupeptin, bestatin and aprotinin were
from Roche, horse heart cytochrome c, arachidonic acid,
hydroxyapatite, phenylmethanesulfonyl fluoride (PMSF)
and pepstatin were from Sigma. DEAE-cellulose was from
Whatman. Ampholines pH 3–10 were from Pharmacia.
Biological preparations
A particulate fraction enriched in plasma membranes was
prepared by centrifugation on a discontinuous sucrose
gradient of a sonicated homogenate of bovine neutrophils in
NaCl/P
i
[12]. The saline buffer (2.7 m
M
KCl, 136.7 m

assuming the presence of two hemes per flavocytochrome b
[17]. The recombinant cytosolic proteins, p47phox, p67phox
and Rac2, prepared as described [18] were kindly provided
by M. C. Dagher UMR 5092 CEA-Grenoble.
The heterodimer S100A8/A9 was purified from bovine
neutrophil cytosol, as described previously [15]. The stained
gel following SDS/PAGE did not show visible traces of
protein contaminant [15]. The identity of S100A8 was further
ascertained by amino-acid sequencing, using Edman degra-
dation. As S100A9 has a blocked N-terminal amino acid,
analysis of the protein was carried out by mass spectrometry.
The protein band corresponding to S100A9 in the gel,
following SDS/PAGE, was excised, and washed 3 times
successively by 25 m
M
ammonium bicarbonate pH 8.0, and
50% acetronitrile/50% 25 m
M
ammonium bicarbonate,
pH 8.0. A final wash with pure water was performed before
complete dehydratation in a vacuum dryer. In gel Ôtryptic
digestionÕ was performed for 4 h at 37 °Cin10 lLof25 m
M
ammonium bicarbonate pH 8.0 with 0.5 lgoftrypsin.A
sample (0.5 lL) of the digestion supernatant was spotted
onto the MALDI sample probe on top of a dried 0.5 lL
mixture of 4 vol. solution of saturated a-cyano-4-hydroxy-
transcinnamic acid in acetone and 3 vol. of nitrocellulose
dissolved in acetone/isopropanol 1 : 1 (v/v). Dried samples
were rinsed by placing a 5-lL vol. of 0.1% trifluoroacetic

antibody was able to interact with the nondissociated
heterodimeric protein S100A8/A9. S100A9 antibodies were
purified from the antiserum by sodium sulfate fractionation.
Other antisera directed against p67phox and p47phox were
obtained from M C. Dagher. Rac1 and Rac2 antibodies
were obtained from Santa Cruz. The phox proteins and the
S100A8/A9 heterodimer were immunodetected after
incubation with goat anti-(rabbit IgG) Ig coupled to
peroxidase. The bound peroxidase was revealed by a
luminescence method using the ECL kit from Amersham.
Protein concentration was assayed with the bicinchonic acid
reagent (BCA) (Bio-Rad) using bovine serum albumin as
standard. Arachidonic acid was dissolved in ethanol and
stored as a stock solution, at a concentration of 200 m
M
.
Protein fractionation of bovine neutrophil cytosol
Protein fractionation was performed on cytosolic extracts of
nonactivated bovine neutrophils obtained by centrifugation
at 140 000 g for 1 h of sonicated homogenates of bovine
neutrophils in NaCl/P
i
supplemented with antiproteases.
Crude cytosol (100 mg protein) was chromatographed
successively on a hydroxyapatite column (7 cm · 2cm)
equilibrated in 20 m
M
Hepes pH 7.5, and on a DEAE
cellulose column (20 · 1 cm) equilibrated in the same
buffer. Elution from the first column (2.4 mL fraction)

and S100A9.
Assay of NADPH oxidase activity after
oxidase activation
The dormant NADPH oxidase of neutrophil membranes
was activated by mixing neutrophil plasma membranes and
the recombinant cytosolic phox proteins, p67phox,
p47phox, GTPcS-loaded Rac2, MgSO
4
and an optimal
amount of arachidonic acid determined for each assay of
oxidase activation [12]. The rate of O
2
–
production by the
activated NADPH oxidase was calculated from the rate of
the superoxide dismutase-inhibitable reduction of ferricy-
tochrome c (100 l
M
)inNaCl/P
i
supplemented with 300 l
M
NADPH at 20 °C. More than 98% of cytochrome c
reduction was sensitive to superoxide dismutase. NADPH
oxidase activity was also assayed by polarographic meas-
urement of the rate of O
2
uptake at 20 °CwithaClark
electrode at a voltage of 0.8 V. All experiments were carried
out at least twice.

amount of p47phox came off by washing with Hepes.
Elution of the hydroxylapatite column by the phosphate
gradient yielded two distinct pools of proteins of interest.
The first one (HTP I fractions 14–17) contained the bulk of
S100A8/A9 and a small, but significant portion of p67phox,
p47phox and Rac. The S100A8/A9 was in large excess with
respect to the cytosolic phox proteins, the disproportion in
these two categories of proteins reflecting probably their
disproportion in crude neutrophil cytosol. In fact S100A8/
A9 represents 10% to 20% of the cytosolic protein content
of bovine neutrophil [13], compared to less than 0.5% for
p47phox and p67phox [19]. The second pool (HTP II
fractions 18–21) consisted of the remainder of S100A8/A9,
accompanied by a majority of p67phox, p47phox and Rac.
Analysis of the protein distribution in the two HTP pools by
SDS/PAGE followed by Coomassie blue staining revealed a
discrete number of proteins, including S100A8 and S100A9
proteins (insert, Fig. 2A). In the HTP I pool, the two
components of the S100A8/A9 complex appeared to be
present in nearly equal amounts, based on SDS/PAGE and
staining by Coomassie blue. In contrast, in the HTP II pool
the relative amount of the S100A9 protein still detectable by
immunodot blot was noticeably lower than that of S100A8
suggesting dissociation of the heterodimer S100A8/A9. Rac
was uniformly distributed in the two HTP pools in contrast
to p67phox and p47phox that were more concentrated in
the second HTP pool than in the first one.
The two HTP pools were further processed separately.
After dilution with Hepes buffer, the proteins of the HTP I
pool were subjected to chromatography on DEAE cellulose

3248 J. Doussiere et al. (Eur. J. Biochem. 269) Ó FEBS 2002
by isoelectric focusing. Aliquots of proteins not retained on
DEAE cellulose (NR), fractions 17–20 (DEAEI) and
fractions 23–25 (DEAEII) were subjected to SDS/PAGE.
Coomassie blue staining of the gel shows an enrichment
of the DEAEI fraction in the two components of the
S100A8/A9 protein with molecular masses of 7 kDa and
23 kDa and the disappearance of a 42–43 kDa protein that
was recovered in the DEAEII fraction and was most likely
actin (insert, Fig. 2B).
Isoelectric focusing of the DEAEI cellulose fraction was
carried out in a 5–40% sucrose gradient supplemented with
ampholines pH 3–10 (Fig. 2C). The bulk of S100A8/A9
and also that of p67phox focused between pH 7.7 and 6.2
(fractions 18–24). These fractions contained only a minor
amount of p47phox. The protein pattern was characterized
by a major peak (FII) with a mean pI value of 6.7 (fractions
21–24), and a shoulder (FI) (fractions 18–20) with a mean pI
value of 7.4, where Rac was concentrated. It is noteworthy
that the pI values of p67phox and p47phox deduced from
the isoelectrofocusing experiment are significantly different
from the theoritical pI values calculated for the isolated
protein, namely 5.9 for p67phox and 9.1 for p47phox. This
is readily explained by the association of these proteins in a
complex. As in the preceding fractionations, eluted proteins
were analyzed by SDS/PAGE, followed by Coomassie blue
staining. At this stage of the protein fractionation the two
components of the S100 protein, S100A8 and S100A9
appeared to be the two major proteins (insert a, Fig. 2C).
Separate immunodot blots carried out with anti Rac1 and

with p67phox and Rac (DEAE III fraction) but not with
p47phox, was eluted with an NaCl gradient between 0.20
M
and 0.38
M
NaCl. Analysis by SDS/PAGE followed by
Coomassie blue staining (insert Fig. 2D) showed that the
NR fraction was enriched in S100A8 accompanied by traces
of S100A9, still immunodetectable by dot blot, whereas the
DEAE III fraction contained the two components of the
heterodimer S100A8/A9. The DEAE III fraction was
characterized by an enrichment in S100A8/A9 and p67, like
the DEAE I fraction (panel B); it was not further processed.
The NR fraction was subjected to isoelectric focusing
(Fig. 2E). About half of p47phox (Fraction FIII) comigrat-
ed with the S100 protein and focused between pH 7.3 and
pH 6.2, i.e. in a zone of pH quite remote from the highly
basic pI of free p47phox. The S100 protein present in
fraction FIII consisted essentially of the S100A8 component
as shown by SDS/PAGE (insert, Fig. 2E) with traces of
S100A9 revealed by immunoblot. Rac was not detectable in
this fraction. The remainder of p47phox free of other
proteins focused at pH of about 9.5, close to the theoritical
pI value, 9.1, of the isolated p47phox in accordance with
the large excess of basic residues in the molecule.
The three-step fractionation described above led to the
following conclusions. 1. Among the cytosolic factors of
oxidase activation, p67phox in association with Rac exhibits
the higher propensity to interact with the heterodimeric
S100A8/A9 protein as shown by their comigration from

90 pmol), p47phox (10 pmol) and GTPcS-preloaded Rac2 (10 pmol) in 40 lLNaCl/P
i
were incubated for 10 min at 20 °C with neutrophil
membranes (4 lg protein corresponding to 1 pmol of heme b). Each well contained a different amount of arachidonic acid ranging from 0 to
7 lmolÆmg membrane protein
)1
. Oxidase activity was initiated by addition of NADPH and cytochrome c in NaCl/P
i
(200 lL) at the final
concentrations of 300 l
M
and 100 l
M
, respectively. Cytochrome c reduction was followed at 550 nm and recorded for 3 min. A complementary
experiment carried out in the presence of 10 lg of SOD showed that more than 98% of the reduction of cytochrome c was inhibited by SOD,
therefore corresponding to the production of O
2
–
. The oxidase activity was expressed in mol of O
2
–
generated per s and per mol of membrane-bound
flavocytochrome b. The traces correspond to the different amounts of p67phox present in the activation medium: j 2pmol,
5pmol,m 10 pmol,
20 pmol, d 40 pmol, n 70 pmol, + 90 pmol. (B) Effect of the presence of S100A8/A9 in the activation medium on the elicitated oxidase activity.
Same conditions as in (A). Curve d corresponds to the control in the absence of S100A8/A9 (experiment shown in (A) carried out with 40 pmol of
p67phox). Curve
shows the effect of S100A8/A9 (64 pmol) added in the activation medium.
3250 J. Doussiere et al. (Eur. J. Biochem. 269) Ó FEBS 2002
of S100A8/A9, the elicited oxidase activity attained a

–
per s per mol heme b
(Fig. 3A). In addition, the shape of the activity curve as a
function of the concentration in arachidonic acid was
different when S100A8/A9 was present in the activation
medium. In the presence of S100A8/A9, a well defined peak
of activity could be observed for a concentration of
arachidonic acid of 1.2 lmolÆmg membrane protein
)1
.This
suggested that S100A8/A9 was able to overcome a limita-
tion in the full expression of oxidase activation, possibly
through the control of an appropriate organization of the
cytosolic factors favoring their productive interaction with
the membrane-bound flavocytochrome b. As S100A8/A9 is
aCa
2+
binding protein, the effect of 1 m
M
Ca
2+
was tested.
No modification of the enhancement of oxidase activation
by S100A8/A9 was observed. It is possible that, due to its
high affinity for Ca
2+
, S100A8/A9 had a full charge of
bound Ca
2+
.

maintained at a fixed value, and the molar ratio of the
cytosolic phox protein (p67phox taken as reference) to
flavocytochrome b was varied up to 40 (Fig. 5). Optimal
arachidonic acid concentration was carefully determined for
all experimental conditions. Inspection of the direct plots of
the elicited oxidase activity indicated an enhancing effect of
Fig. 4. Kinetic parameters of elicited NADPH oxidase after activation in the presence or absence of S100A8/A9. Membranes from bovine neutrophils
(290 lg protein equivalent to 110 pmoles of flavocytochrome b) were incubated at 20 °C with p67phox (1480 pmol), p47phox (370 pmol), GTPcS-
loaded Rac2 (370 pmol) and arachidonic acid (1.2 lmol/mg membrane protein) in a volume of 450 lL(d). A parallel incubation was carried out in
the presence of 2500 pmol of S100A8/A9 (s). Following incubation, 10 lg protein aliquots were withdrawn for measurement of O
2
–
generation by
the superoxide dimutase inhibitable reduction of cytochrome c (A), and 100 lg protein aliquots for measurement of O
2
uptake with a Clark
electrode (B). The rate of O
2
–
production expressed in lmol generated per min and per mg of membrane protein was measured in a spectro-
photometric cuvette filled with 2 mL of NaCl/P
i
supplemented with 100 l
M
cytochrome c and different concentrations of NADPH. In the case of
O
2
uptake, the O
2
concentration of the medium was lowered to 80–90 l

it ensues that, at a molar ratio of S100A8/A9 to p67phox of
about 2, the kinetics of the elicited oxidase were virtually
linear. In brief, in the absence of S100A8/A9, nonmichaelian
kinetics were observed for the rate of production of O
2
–
by the
membrane-bound flavocytochrome b activated by increas-
ing amounts of the phox cytosolic proteins. Addition of
S100A8/A9 renders the kinetics michaelian. The fact that
michaelian kinetics are attained, using a ratio of S100A8/A9
to p67phox as low as 2, suggests interaction between the
two proteins within a complex.
Effect of S100A8/A9 on the time course
of oxidase activation
NADPH oxidase activation in a cell-free system is a process
which at room temperature reaches its maximum after
several minutes at 20 °C. In the following experiment, we
determined the effect of S100A8/A9 on the time course of
oxidase activation. For technical convenience, the time
course of oxidase activation was analyzed by measuring the
elicited oxidase activity of membrane-bound flavocyto-
chrome b in terms of O
2
uptake in an oxygraphic cell. The
cytosolic phox proteins (p67phox, p47phox and GTPcS-
loaded Rac2 present in a molar ratio of 3/1/1) were left in
contact for 5 min at 20 °C with or without the S100A8/A9
protein (molar ratio of S100A8/A9 to p67phox adjusted to a
value of 3.3). It should be noted that, like in the experiment

uptake depends directly on
the concentration of Fba. It followed that the oxidase
activity (v) should reach a maximal value (V) when the
totality of the flavocytochrome (Fbt) is activated. On
the basis of this hypothesis, the first order equation
v ¼ V (1–e
–kt
) was used to describe the rate of flavocyto-
chrome b activation. In the case of preincubation of the
cytosolic phox proteins with S100A8/A9, the experimental
data fitted well with the theoretical curve derived from the
above first order reaction, yielding a V-value of 300 nmol
O
2
uptake per min and per mg of membrane protein, and a
rate constant k of 0.012 s
)1
(Thick line in Fig. 6). In
contrast, in the absence of S100A8/A9, the experimental
curve could not be fitted with a single first order reaction
curve (thin line in Fig. 6). However, a good fit was found by
using the sum of two first order equations, the first one
being characterized by a k-value of 0.112 s
)1
and a V of
110 nmol O
2
per min and per mg of membrane protein, and
the second by a k-value of 0.008 s
)1

data).
3252 J. Doussiere et al. (Eur. J. Biochem. 269) Ó FEBS 2002
It is noteworthy that the sum of these two rates, i.e.
221 nmol O
2
per min and per mg of membranes protein is
lower than the rate of O
2
uptake measured in the presence of
S100A8/A9, namely 300 nmol O
2
per min and per mg of
membrane protein. These results led us to hypothesize that
in the absence of S100A8/A9 the cytosolic phox proteins are
organized in at least two pools probably in slow equilib-
rium, one of which is much more efficient than the other in
oxidase activation. The effect of S100A8/A9 would be to
bind and rearrange the totality of the cytosolic phox
proteins in a complex capable of activating optimally
flavocytochrome b. Association of S100A8/A9 with the
cytosolic phox proteins might however, bring some steric
constraint, so that oxidase activation during the first minute
proceeds more slowly in the presence of S100A8/A9 than its
absence (Fig. 6). As the maximal oxidase activity measured
in the presence of S100A8/A9 was significantly greater than
in its absence, we can tentatively conclude that S100A8/A9
enhances the recruitment of the cytosolic phox proteins to
the membrane-bound flavocytochrome b and (or) stabilizes
their interaction.
DISCUSSION

solic phox proteins appear not to interact productively any
more with their target sites on flavocytochrome b in the
absence of S100A8/A9. This observation corroborates the
fact that, in absence of S100A8/A9, the time course of
oxidase activation may be fitted by the sum of two first order
kinetics (Fig. 6). When the cell-free medium was supplemen-
ted with S100A8/A9, michaelian kinetics were restored and
an homogeneous first order reaction was found for the
production of O
2
–
. Through its association with cytosolic
phox proteins, the S100A8/A9 protein might act as a scaffold
protein, that favors the organization of the phox proteins in a
reactive complex and helps this complex to interact in a
productive manner with flavocytochrome b to activate its
dormant oxidase activity. It might also simply favor the
delivery of bound arachidonic acid to the oxidase complex.
The S100A8/A9 heterodimer specifically binds long-chain
unsaturated fatty acids and, most particularly, arachidonic
Fig. 6. Effect of S100A8/A9 on the time course of oxidase activation.
Oxidase activity was measured by the rate of O
2
uptake with a Clark
electrode (cf. Materials and methods). The cytosolic factors (CF)
p67phox (420 pmol), p47phox (140 pmol) and GTPcS-preloaded
Rac2 (140 pmol) were incubated together with 2.5 m
M
MgSO
4

was however, possible by combining two first order reaction curves
(insert B), from which maximal rates and rate constants were calcu-
lated (see text for details).
Ó FEBS 2002 S100A8/A9 is a partner of p67phox in neutrophils (Eur. J. Biochem. 269) 3253
acid in a calcium-dependent manner, whereas the S100A8
or S100A9 homodimers lack this property [6–8]. Data
obtained through different experimental approaches suggest
that arachidonic acid is an in vivo activator of the NADPH
oxidase. Through the use of a model of cytosolic phos-
pholipase A2 (cPLA2) deficient phagocyte-like cells, it was
demonstrated that cPLA2-generated arachidonic acid is
essential for the activation of NADPH oxidase [22]. It was
also reported that the large subunit of flavocytochrome b,
gp91phox, constitutes an arachidonic acid-activated proton
channel (reviewed in [23,24]). In addition, it is noteworthy
that arachidonic acid at low concentration induces the
transition of the heme iron of flavocytochrome b from an
inactive hexacoordinated form to a pentacoordinated form
capable of binding O
2
[25]. In a previous study on oxidase
activation carried with neutrophil membranes and crude
cytosol [12], we found that in the presence of relatively high
concentrations of arachidonic acid, the increase in the
turnover of NADPH oxidase depended on the amount of
the cytosolic fraction present in the cell-free system, in other
words of the amount of cytosolic phox proteins capable of
binding to flavocytochrome b. The S100A8/A9 present in
crude cytosol and artificially loaded with arachidonic acid
was probably beneficial to this process. Because of its high

the two types of cells might be responsible for this apparent
lack of correlation.
Components of the neutrophil cytoskeleton, namely actin
and coronin have been reported to interact with the
cytosolic phox proteins, the b actin with p47phox [27] and
coronin with p40phox [28]. Most significantly, abnormali-
ties in O
2
–
production have been found in neutrophils of a
patient carrying a mutation in nonmuscle actin [29]. These
data and ours on S100A8/A9 strongly suggest that there
exists a complex array of interactions between the classical
phox components of the oxidase complex and other
molecular protein species in phagocytic cells. The function
of these ancillary species might be to regulate the kinetics of
the production of O
2
–
in the respiratory burst and to
segregate activated oxidase complexes in the phagosomal
membrane.
ACKNOWLEDGEMENTS
We thank Dr J. Willison for careful reading of the manuscript and
Mrs Bournet-Cauci for excellent secretarial assistance, Dr Marie-Claire
Dagher for the gift of recombinant cytosolic phox proteins, Drs Anne-
Christine Dianoux and Marie-Jose
´
Stasia for anti S100A9 antibodies,
and Dr Je

Holtkemper,U.&Sorg,C.(1997)Theheterodimerofthe
Ca
2+
-binding protein MRP-8 and MRP-14 binds arachidonic
acid. FEBS Lett. 408, 81–84.
7. Siegenthaler, G., Roulin, K., Chatellard-Gruaz, D., Holtz, R.,
Saurat, J.H., Hellman, U. & Hagens, G. (1997) A heterocomplex
formed by the calcium-binding proteins MRP8 (S100–A8) and
MRP-14 (S100A9) binds unsaturated fatty acids with high affinity.
J. Biol. Chem. 272, 9371–9377.
8. Kerkhoff, C., Klempt, M., Kaever, V. & Sorg, C. (1999) Two
calcium-binding proteins, S100A8 and S100A9, are involved in the
metabolism of arachidonic acid in human neutrophils. J. Biol.
Chem. 274, 32672–32679.
9. Bromberg, Y. & Pick, E. (1984) Unsaturated fatty acids stimulate
NADPH-dependent superoxide production by cell-free system
derived from macrophages. Cell Immunol. 88, 213–221.
10. De Leo, F.R. & Quinn, M.T. (1996) Assembly of the phagocyte
NADPH oxidase. Molecular interaction of oxidase proteins.
J. Leukoc. Biol. 60, 677–691.
11.Wientjes,F.B.,Hsuan,J.J.,Totty,N.F.&Segal,A.W.(1993)
P40phox, a third cytosolic component of the activation complex of
the NADPH oxidase to contain src homology domain 3. Biochem.
J. 317, 919–924.
12. Doussie
`
re, J., Bouzidi, F., Poinas, A., Gaillard, J. & Vignais, P.V.
(1999) Kinetic study of the interaction of the neutrophil NADPH
oxidase by arachidonic acid. Antagonistic effects of arachidonic
acid and phenylarsine oxide. Biochemistry 38, 16394–16406.

2
generating NADPH oxidase in a semi-
recombinant cell-free system. Assessment of the function of Rac in
the activation process. Eur. J. Biochem. 226, 587–595.
19.Jouan,A.,Pilloud-Dagher,M C.,Fuchs,A.&Vignais,P.V.
(1993) A generally applicable ELISA for the detection and
quantitation of the cytosolic factors of NADPH oxidase activa-
tion in neutrophils. Anal. Biochem. 214, 252–259.
20. Freeman, J.L. & Lambeth, J.D. (1996) NADPH oxidase activity is
independent of p47phox in vitro. J. Biol. Chem. 271, 22578–22582.
21. Koskhin, V., Lotan, O. & Pick, E. (1996) The cytosolic component
p47phox is not a sine quanon participant in the activation of
NADPH oxidase, but is required for optimal superoxide pro-
duction. J. Biol. Chem. 271, 30326–30329.
22. Dana, R., Leto, T., Malech, H. & Levy, R. (1998) Essential
requirement of cytosolic phospholipase A2 for activation of the
NADPH oxidase. J. Biol. Chem. 273, 441–445.
23. Henderson, L.M., Banting, G. & Chappell, J.B. (1995) The ara-
chidonic-activable NADPH oxidase-associated H
+
channel. Evi-
dence that gp91phox functions as an essential part of the channel.
J. Biol. Chem. 270, 5909–5916.
24. Henderson, L.M. & Chappell, J.B. (1996) NADPH oxidase in
neutrophils. Biochim. Biophys. Acta 1273, 87–107.
25. Doussie
`
re, J., Gaillard, J. & Vignais, P.V. (1996) Electron transfer
across the O
À